> ## Documentation Index
> Fetch the complete documentation index at: https://safety.flamy.net/llms.txt
> Use this file to discover all available pages before exploring further.

# Safe filters

> How Flamify Safe Mode filters NSFW catalog, media, and generation for PG-13 and app-store brand surfaces.

# Safe filters

> Brand-aware Safe Mode for PG-13 and app-store surfaces — without removing
> the adult product where it is intended.

Flamify operates multiple brand fronts. Some must remain suitable for app
stores and general audiences. **Safe Mode** (internally: `APP_PALM`) is the
per-brand switch that turns on SFW filtering for those surfaces.

Safe Mode is **not** the same as the chat forbidden-topic pipeline. Forbidden
topics apply platform-wide to user messages. Safe Mode additionally reshapes
**what is discoverable and generatable** on a given brand.

## When Safe Mode is on

For a brand with Safe Mode enabled (resolved from the active brand header):

### Catalog and discovery

* Feed, explore, tags, and SEO category pages return only characters marked
  **`is_safe = true`**.
* Cache keys are segmented by Safe Mode so an NSFW tag list can never leak
  into a Safe Mode response.
* Character profile pages for non-safe characters return not found on Safe
  Mode surfaces.

### Media visibility

On Safe Mode brand surfaces, character media is limited to mild tiers
(profile / live / background / mild). Suggestive, bold, naughty, and
forbidden tiers are hidden. Purchasing / unlocking restricted tiers is
blocked.

### Chat and generation

* Outbound AI requests carry `is_safemode = true`.
* Chat is configured with the Safe Mode model / parameter profile.
* Media prompt building:

  * Strips NSFW vocabulary from user free-text.
  * Rejects explicit scene types.
  * Adds NSFW terms to the negative prompt so the image model is steered away
    from adult output.

### What users see

A PG-13 catalog of operator-marked safe characters, mild media only, and
generation constrained to SFW output. The adult Flamify product remains
available on brands where Safe Mode is off.

## Character `is_safe` flag

`is_safe` is an explicit operator / admin flag meaning “safe for app-store
review mode.” It is not automatically derived from a single generative pass.

Operators are expected to mark only characters that are appropriate for Safe
Mode distribution. Characters that are adult-oriented stay available on the
main product but are excluded from Safe Mode catalogs.

## Always-on age blockers (independent of Safe Mode)

Even when Safe Mode is off, appearance / image negative prompts include
**child / underage / teen** blockers. Adult erotic content may be allowed on
the main product; content implying minors is not.

## Relationship to prohibited content

Safe Mode does **not** relax [Prohibited content](/policies/prohibited-content).
Minors, CSAM signals, self-harm facilitation, and other hard limits remain
enforced via the chat moderation pipeline on every brand.

Safe Mode only adds an extra distribution and generation constraint for
brands that must stay SFW.

## Summary for payment / compliance reviewers

| Control                   | Scope                  | Effect                                                            |
| ------------------------- | ---------------------- | ----------------------------------------------------------------- |
| Forbidden-topic STOP/HOLD | All brands, chat input | Blocks or warns on prohibited topics                              |
| Character age ≥ 18        | All brands             | Schema rejects underage characters                                |
| Appearance age negatives  | All brands, media      | Steers models away from underage visuals                          |
| Safe Mode catalog filter  | Brands with `APP_PALM` | Only `is_safe` characters                                         |
| Safe Mode media tiers     | Brands with `APP_PALM` | Hides NSFW media; blocks unlock                                   |
| Safe Mode generation      | Brands with `APP_PALM` | Strips NSFW prompts; blocks explicit scenes; NSFW negative prompt |

Together, these controls let Flamify offer an adult companion product while
keeping designated brand surfaces suitable for stricter distribution
channels.